MARINE ECOLOGY PROGRESS SERIES Mar Ecol Prog Ser

Vol. 350: 127–136, 2007 doi: 10.3354/meps07128

Published November 22

Foraging behaviour and energetics of Cape gannets Morus capensis feeding on live prey and fishery discards in the Benguela upwelling system Lorien Pichegru1, 2,*, Peter G. Ryan2, Carl D. van der Lingen3, Janet Coetzee3, Yan Ropert-Coudert 4, David Grémillet1, 2 1

Centre National de la Recherche Scientifique, DEPE-IPHC, 23 rue Becquerel, 67087 Strasbourg, France DST/NRF Centre of Excellence at the Percy FitzPatrick Institute of African Ornithology, University of Cape Town, Rondebosch 7701, South Africa 3 Marine and Coastal Management, Department of Environmental Affairs and Tourism, Private Bag X2, Rogge Bay 8012, South Africa 4 National Institute of Polar Research, 1-9-10, Kaga, Itabashi-ku, Tokyo 173-8515, Japan 2

ABSTRACT: We compared the foraging ecology of Cape gannets Morus capensis attending 2 colonies of equivalent size, yet with contrasting diet and population trends. One colony, on the west coast of South Africa, is decreasing in size and its occupants feed mainly on fishery wastes, whereas the other colony, on the south coast of South Africa, is growing and its occupants feed exclusively on natural prey (pelagic fish). In October and November 2005, we examined the diet, at-sea behaviour, and energy requirements of breeding gannets using direct observations, miniaturised GPS loggers, and time-depth recorders attached to foraging adults. Concurrent hydroacoustic surveys allowed us to assess the distribution and abundance of their preferred prey (the sardine Sardinops sagax and anchovy Engraulis encrasicolus). Birds from the declining west coast colony foraged in areas containing very low abundances of pelagic fish. They fed primarily on low-energy fishery discards. They increased their foraging effort and exploited a greater area than birds from the growing colony, which took advantage of abundant pelagic fish stocks in their foraging range. A marked eastward shift of pelagic fish initiated in the late 1990s has resulted in the shortage of natural prey to Cape gannets on the west coast, strongly suggesting that the local population trend is driven by food availability during the breeding season. A bioenergetic model showed that enhanced availability of lowenergy fishery discards does not seem to compensate for the absence of natural prey. Added to the predation pressure by the Cape fur seal Arctocephalus pusillus and the great white pelican Pelecanus onocrotalus, those threats weigh heavily on a vulnerable seabird population. KEY WORDS: Bioenergetic modelling · Fishery waste · GPS tracking · Hydroacoustic survey · Marine management · Pelagic fish · Seabird Resale or republication not permitted without written consent of the publisher

To survive and reproduce, adult seabirds have to gather food at sea in a highly variable environment (Bakun 1996). Their ability to adapt to this variability and forage efficiently determines their population trajectories (Lewis et al. 2006). On an evolutionary time scale, seabird life-history traits are adapted to cope with environmental stochasticity through slow chick

growth rates, long breeding seasons, delayed maturity, and long lifespans. On a shorter time scale, as during the breeding season, adult seabirds show flexible foraging behaviour, which should help them cope with labile prey distributions (e.g. Lescroël & Bost 2005). Studies of seabird at-sea behaviour are currently booming thanks to rapid advances in biotelemetry technology. However, very few field studies have simultaneously investigated the foraging behaviour

*Email: [email protected]

© Inter-Research 2007 · www.int-res.com

INTRODUCTION

128

Mar Ecol Prog Ser 350: 127–136, 2007

and predatory efficiency of seabirds relative to the distribution and abundance of their prey (Grémillet et al. 2004a). This is because seabirds often forage over huge oceanic areas, and population studies of their prey (mainly fish) require at-sea investigations on board research vessels, which are both expensive and time-consuming. Beyond natural environmental variability, seabirds also face the impact of human activities, of which the most profound are probably industrial fisheries. These fisheries often deplete marine fish stocks, diminishing the availability of food for numerous seabird species (Furness & Tasker 2000). Conversely, they may also generate large volumes of fish waste, upon which many seabird species have learnt to make a living (Garthe 1996). Several studies have investigated the impact of fishery discards on seabird behaviour, ecology, and population dynamics (Arcos & Oro 1996, Oro et al. 2004), but no study has compared the foraging behaviour of a seabird species feeding either on its natural live prey or on fishery discards and discussed the consequences of these different foraging options on population processes. We performed such a study on the Cape gannet Morus capensis, a large coastal seabird endemic to the Benguela upwelling system; the birds primarily feed by plunge diving on small pelagic fish (Berruti et al. 1993). The Benguela, characterised by high productivity due to the vertical transport of cold, nutrient-rich waters to the surface, extends from southern Angola (17° S) to the southern tip of Africa (34° S) and eastwards to East London, South Africa (28° E). It is one of the most productive ocean areas in the world (Shannon & O’Toole 2003), but is also heavily exploited by fisheries that target small pelagic fish, such as the sardine Sardinops sagax and the anchovy Engraulis encrasicolus, as well as demersal whitefish such as the hake Merluccius spp. (Crawford et al. 1987). Fisheries compete with seabirds by depleting stocks of their natural prey (Crawford 1999), but also produce large amounts of waste, some of which is consumed by seabirds. In the 1980s, hake fishery discards represented 15% of seabird consumption in the Benguela (i.e. 65 000 t yr–1, Crawford et al. 1991). Towards the end of the 1990s, approx. 9000 t of hake were still discarded annually off the west and south coasts of South Africa (Walmsley et al. 2007). The Cape gannet breeds at 6 island colonies from central Namibia to the southeast coast of South Africa (Fig. 1). We studied the foraging ecology of birds from the 2 southernmost colonies located on opposite sides of Cape Agulhas: Malgas Island on the west coast of South Africa and Bird Island on the south coast (Fig. 1). These colonies were of equivalent size at the time of our study (approx. 65 000 pairs on each island), together supporting 80% of the world breeding popu-

lation of Cape gannets (Crawford 2005). However, they have experienced different population trends: the population on Bird Island has shown a positive growth rate since the start of monitoring in the 1950s (Klages et al. 1992), whereas that on Malgas Island grew more slowly until 1996 (Crawford 1999) and has recently declined (Crawford et al. 2007). The colonies grew thanks to the progressive cessation of guano activities and reduction of human disturbance (Griffiths et al. 2004). Studies of the diet of Cape gannets were conducted on both islands in the 1980s. Birds from Bird Island fed exclusively on natural prey (i.e. sardines, anchovies, and saury Scomberesox saurus) (Klages et al. 1992), whereas birds from Malgas fed on both live prey and fishery discards (Berruti et al. 1993). Because a recent study suggested that Cape gannet population trends are driven by food availability during the breeding season (Lewis et al. 2006), we investigated the foraging strategies of Cape gannets from Malgas Island and Bird Island to test the hypothesis that contrasting food abundance and quality (live prey versus fishery waste) contribute to the contrasting population trends at the 2 different colonies. We compared the foraging behaviour and diet of gannets from each island with the abundance and distribution of their principal prey (small pelagic fish: the sardine Sardinops sagax and the anchovy Engraulis encrasicolus) as determined during an acoustic survey conducted concurrently with our studies on the birds. We then estimated the energetic needs of Cape gannet for breeding successfully on each island using a bioenergetic model.

MATERIALS AND METHODS We studied the foraging behaviour of Cape gannets breeding at Malgas Island (Saldanha Bay, 33° 03’ S, 17° 55’ E) in the Western Cape and at Bird Island (Nelson Mandela Bay, 33° 50’ S, 26° 17’ E) in the Eastern Cape of South Africa (Fig. 1). Observations were conducted from 9 to 31 October 2005 on Malgas Island and from 21 to 29 November 2005 on Bird Island, under permits issued by South African National Parks. Cape gannets on Bird Island typically breed 1 to 2 months later than those on Malgas Island, and therefore birds from the 2 colonies were at the same breeding stage. Recording seabird foraging behaviour. On each island, 30 adult birds raising small chicks (1 to 3 wk old) were caught on their nest site prior to a foraging trip and fitted with electronic devices. The birds were caught when both partners were at the nest site, so that 1 partner could stay on the nest to guard the chick. Birds were caught with a hook mounted on a telescopic pole, which allowed us to capture birds nesting away from the margins of the colony.

129

Pichegru et al.: Seabird foraging on live prey versus fishery discards

S 28° 6° 12°

ay

ip B

kl nde

South Africa

Ho

30° 0m

y Ba e bin lum

t’s ber

30° 36°

Lam Co ape

Malgas Island

10° 16° 22° 28°

Cape Point

Bird Island

NW Flow

36°

38°

14°

34° 40°

LonEast Ma Nels don nde on C n l aB Tow y s ay a a s B i e y e h c p p Ba Agul fanta ossel Ca Cat-Fran se n M S Fal Cape ape I C s die Ed

Jet Current

34°

24°

N

20

32°

18°

Currents 16°

18°

20°

22°

24°

26°

28°E

Fig. 1. Morus capensis. Distribution of 6 colonies (inset) and main ocean currents in Southern Africa (modified from Miller et al. 2006)

Three types of data logger were used: (1) GPS-TD loggers (earth & Ocean Technologies) are miniaturised GPS devices combined with timedepth recorders (see Ryan et al. 2004 for details). Latitude and longitude were recorded at 2 min intervals to an accuracy of approximately 10 m, and depth and temperature were recorded at 1 s intervals to the nearest 0.1 m. Loggers were housed in waterproof and pressure-tight, streamlined fibre-composite containers closed with an O-ring cap (96 × 39 × 26.5 mm; mass 75 g, i.e. 2.4% of adult body mass). (2) GPS data loggers (Technosmart) were also housed in the same type of waterproof and pressuretight, streamlined containers as above (95 × 48 × 24 mm; mass 65 g, i.e. 2.1% adult body mass; see Grémillet et al. 2004b for details). Position data (latitude, longitude, and altitude) were recorded at 10 s intervals, with an accuracy of 10 m. (3) Time-depth recorders (TDR, M190-D2GT, 12-bit resolution, 60 × 15 mm, 15 g; Little Leonardo) were deployed in conjunction with the GPS loggers. These devices monitor depth and temperature every second with an absolute accuracy of 0.1 m (see Ropert-Coudert et al. 2004 for details). The GPS + TDR package weighed 80 g, i.e. 2.5% of adult body mass. All birds were equipped for a single foraging trip to minimise pseudo-replication problems. The GPS device was attached to the bird’s lower back, using 3 central tail feathers, while TDRs were attached under-

neath the tail. All devices were fixed with Tesa tape, which allows the loggers to be removed with minimal damage to the plumage. Handling lasted 4 to 7 min from capture to release and took place under shade to avoid heat stress, while the bird’s head was covered to reduce handling stress. Although these techniques have been used on several occasions on Cape and Northern gannets Morus bassanus (Grémillet et al. 2004b, 2006, Lewis et al. 2006) without any apparent impact on the welfare of the animals, we checked for potential biases resulting from human handling or impacts from the presence of loggers. Ten birds from control nests with comparable chick age were marked using biocompatible dye on a paintbrush fixed to a pole. Marking took place without handling the birds. Control and experimental nests were then monitored every hour from dawn to dusk until a complete foraging trip had been performed by each partner. We assumed that the foraging trip of equipped birds would be modified in case of disturbance (Adams & Klages 1999). Control nests were then checked every hour for 4 further complete consecutive foraging trips to assess the variability of the foraging environment through the regularity of trip duration. We assumed that in a predictable environment, birds have more regular foraging trips and regular daily feeding rates of young (Schreiber & Burger 2002). We compared the standard deviations of the mean duration of the 5 foraging trip durations between the 2 colonies.

130

Mar Ecol Prog Ser 350: 127–136, 2007

Analysis of electronic data set and spatial analysis. Dive profiles were analysed using IGOR Pro (v. 4.01, Wavemetrics), with a dive defined as when a device recorded a depth ≥0.5 m for ≥1 s. We determined the number of dives per h of foraging trip, the average dive duration, and the average maximum dive depth for each of the foraging trips (see details in Ropert-Coudert et al. 2004). Using the GPS positions, we determined and compared the foraging range of the birds from the 2 colonies. Foraging range was determined from filtered GPS positions known to be associated with feeding behaviour, i.e. positions where the birds were flying (speed > 10 km h–1) and displaying a sinuous path (see details in Grémillet et al. 2004b). This method discards commuting sections, thus minimising pseudo-replication, and has been validated by Grémillet et al. (2006). Positions obtained from GPS loggers recording at 10 s intervals were sub-sampled to have 1 position every 2 min to be comparable with data obtained from the GPS-TD loggers. Adaptive kernel analyses of bird foraging positions were used to assess the time spent per unit area. Analyses were conducted using Arcview GIS 3.2 with the smoothing factor chosen according to the Least Square Cross Validation (LSCV) method (Girard et al. 2002). Contour levels covering 10 to 90% of the foraging locations were estimated. The surface of the foraging range was calculated through the concave polygon method with Ranges VI (Anatrack). We also calculated the average density of sardines and anchovies present in the foraging area of the birds from both islands. Energetic modelling. We used the algorithm of Enstipp et al. (2006) to assess the daily energy requirements (kJ d–1), daily food intake (g d–1), and metabolic scope (average metabolic rate expressed as a multiple of the basal metabolic rate [BMR]) of Cape gannets. We used time budget information gained from the GPS data loggers, such as the foraging trip duration and the time spent flying, and field data for the body mass and chick age. Birds equipped with loggers were weighed before attaching the devices and again when back at the nest, after removal of the devices, i.e. 1 to 2 h after the bird returned to its nest. This method allowed us to minimise the food loss due to the bird regurgitating during its handling and to standardise the amount of food already fed to the chick (R. Navarro unpubl. data). Adult gannets were weighed to the nearest 25 g using a spring balance (Salter Brecknell Super Samson, RACO Industries). There was no difference in adult body mass before a foraging trip between gannets from the 2 islands (F1,56 = 3.67, p = 0.061). Diet samples of gannets were collected during our study by Marine and Coastal Management staff from 20 to 22 October on Malgas Island and from 14 to

17 November on Bird Island. Prey types and their proportions by mass in the overall diet samples were assessed. We assumed that food brought back to the nest by birds (i.e. the difference in bird mass before and after a foraging trip) would be composed of prey in the same respective proportions as in the overall diet samples. We then calculated the energetic value of the food brought back by each bird according to calorific values from the literature (Batchelor & Ross 1984). Fish distribution and abundance. Sardines and anchovies are the preferred prey of Cape gannets (Adams & Klages 1999) and can constitute > 90% of their diet (Crawford 2005). Information on the abundance and distribution of these small pelagic fish species over the period of observations of gannet foraging was obtained from a hydroacoustic survey conducted by Marine and Coastal Management from 18 October to 5 December 2005. This survey mapped the distribution and estimated the abundance of anchovies, sardines, and the redeye round herring Etrumeus whiteheadi off South Africa between Hondeklip Bay on the west coast and East London on the east coast (Fig. 1). It consisted of transects oriented perpendicular to the coast and extending across the continental shelf from close inshore to the 200 m isobath. Echo-integration techniques were used to estimate fish density along survey transects, and midwater trawling was used to determine the species composition and size frequency distributions of pelagic fish. A full description of the methods is given in Barange et al. (1999). Statistical analyses. The threshold for statistical significance, α, was 5%. To fulfil the criteria of normality, trip duration was calculated excluding the nighttime (from the ephemerid), given that gannets do not fly at night (Ropert-Coudert et al. 2004, pers. obs.). However, in order to be comparable with other studies, nighttime was included in the calculation of the foraging trip duration versus foraging path length regression. To compare variables between the 2 islands, we performed balanced ANOVA, with all the variables as responses and islands as model when the data had a normal distribution (sometimes after square-root transformation), equal variance, and the same number of values. Otherwise, Kruskal-Wallis tests were performed for each variable.

RESULTS Seabird foraging behaviour On Malgas Island, we equipped 21 birds with a single GPS and 9 with a GPS-TD. On Bird Island, we equipped 18 birds with a GPS coupled with a TDR, 5 with a GPS-TD, and 7 with a single GPS. Although

Pichegru et al.: Seabird foraging on live prey versus fishery discards

131

Table 1. Morus capensis. Adult body mass and foraging characteristics. some birds stayed at sea for longer than *p < 0.05, **p < 0.01, ***p < 0.001, ns: not significant, nd: not determined the loggers’ battery capacity, 27 complete GPS tracks were collected from both Parameter Malgas Bird Signifiislands, as well as complete diving activity Island Island cance over a trip for 7 birds from Malgas Island and 14 birds from Bird Island. All data logBody mass (g) 2673 ± 250 2785 ± 190 ns gers were retrieved after a single foraging Trip duration incl. night (h) 29.73 ± 11.1 22.83 ± 10.68 nd Trip duration excl. night (h) 17.4 ± 7.19 13.8 ± 17.18 * trip and all equipped birds continued to Foraging path length (km) 504 ± 237 367 ± 232 * breed normally. No difference in the foragMax. distance to colony (km) 136 ± 64 112 ± 89 ns ing trip duration was found between Mean speed (km h–1) 48.5 ± 3.14 47.09 ± 5.31 ns equipped and control birds at either site Time spent flying (h) 10.1 ± 4.4 6.55 ± 3.4 ** Time spent flying per trip (%) 60 ± 14 49 ± 15 ** (Malgas Island: F1,31 = 0.42, p = 0.523; Bird Food brought to nest (g) 361 ± 243 213 ± 141 * Island: F1,57 = 0.86, p = 0.357). Energy brought back to nest (kJ) 1561 ± 1053 1162 ± 1103 ns Gannets from Malgas Island worked Index of trip regularity 0.71 ± 0.38 0.35 ± 0.15 *** Dives per trip h 2.18 ± 1.3 2.42 ± 2.3 ns harder than those from Bird Island Mean duration of dives (s) 4.58 ± 3.78 5.72 ± 4.38 ns (Table 1). Foraging trips of Malgas Island Mean max. dive depth (m) 3.11 ± 2.48 4.24 ± 2.37 * birds lasted longer (F1,52 = 4.1, p = 0.048) Foraging range (km2) 31 191 6673 and showed a greater foraging path length (F1,52 = 6.06, p = 0.017) compared to trips Island (T1,23 = 5.02, p < 0.001, r2 = 0.231: food brought to by Bird Island gannets, and Malgas birds also spent a greater amount of time flying per trip h (F1,52 = 7.88, nest (g) = 683 – [10.3 × trip duration (h)]) than for birds p = 0.007). They brought a greater amount of food back from Bird Island (T1,23 = 3.06, p = 0.005, r2 = 0.001: food to the nest (H1,50 = 5.62, p = 0.018) whilst diving shalbrought to nest (g) = 203 + [0.41 × trip duration (h)]). lower (F1,52 = 5.84, p = 0.026), but the mean energy Gannets from Bird Island showed less variability in forcontent of their food was not significantly different aging trip duration, i.e. with smaller standard deviafrom that of birds from Bird Island (H1,54 = 0.16, p = tions of the average foraging trip durations (H1,54 = 0.690). The maximum distance travelled from the 17.11, p < 0.001) compared to those from Malgas. colony was significantly correlated with foraging trip duration for Malgas birds (F1,25 = 46.78, p = 0.000, r2 = 0.638: max. distance from colony (km) = 9.7 + [173 × Seabirds and the distribution and availability of their trip duration (h)]), but not for birds from Bird Island pelagic prey (F1,25 = 2.81, p = 0.106, r2 = 0.065). The mass of food brought back to the colony was significantly correlated Raw GPS tracks of the birds from the 2 islands are with the time spent at sea for both islands, but the corshown on Fig. 2. Birds from both colonies covered a relation coefficient was greater for birds from Malgas major axis approx. 500 km long. However, the area

Fig. 2. Morus capensis. GPS tracks of foraging birds. Different colours show tracks from different birds (n = 27 birds from each island)

132

Mar Ecol Prog Ser 350: 127–136, 2007

actively exploited by the foraging birds was 5 times greater at Malgas (31 191 km2) than at Bird Island (6673 km2) (Table 1). The hydroacoustic survey recorded virtually no sardines or anchovies off the west coast, indicating an almost complete absence of natural prey for gannets from Malgas Island, which foraged from offshore of Lambert’s Bay to Cape Agulhas (see Fig. 3). Sardines were distributed from False Bay to Nelson Mandela Bay, with the highest densities (>100 g m–2) located towards the shelf edge around 150 km offshore off Mossel Bay and in Nelson Mandela Bay. Anchovies were distributed over much of the continental shelf between False Bay

and Cape Infanta, towards the shelf edge off Mossel Bay, and inshore near Cape St Francis, with several moderate- to high-density concentrations (50 to >100 g m–2). Spatial analysis on the foraging range of the birds revealed that most gannets from Malgas Island foraged in areas where few natural prey were observed by the hydroacoustic survey, whereas the foraging core of the birds from Bird Island occurred over a concentration of sardines in Nelson Mandela Bay (Fig. 3). The average density of natural prey located in the foraging area of birds from Malgas Island was 7.2 and 18.0 g m–2 for sardines and anchovies, respectively,

a SOUTH AFRICA

East London

Malgas

Bird Island Port Elizabeth

Cape Town Mossel Bay

Foraging Anchovy density core 0.1 0.05–0.1 0.1–0.2 0.1–0.3 0.2–0.3 0.3–0.5 0.3–0.4 0.5–0.7 0.4–0.5 0.7–1 0.5–0.6 1–10 0.6–0.7 0.7–0.8 0.8–0.9

Cape Agulhas

0

200 km

b SOUTH AFRICA

Malgas

East London

Bird Island Port Elizabeth

Cape Town Mossel Bay Cape Agulhas

0

200 km

Sardine Foraging density core 0.1 0.05–0.1 0.1–0.2 0.1–0.3 0.2–0.3 0.3–0.5 0.3–0.4 0.5–0.7 0.4–0.5 0.7–1 0.5–0.6 1–10 0.6–0.7 0.7–0.8 0.8–0.9

Fig. 3. Morus capensis. Foraging range (density of GPS positions associated with feeding behaviour; n = 27 birds from each island) in relation to the distribution and abundance of (a) anchovies and (b) sardines. Colour code for foraging range gives the contour levels covering 10 to 90% of the locations. Fish density is given in percentile

Pichegru et al.: Seabird foraging on live prey versus fishery discards

whereas birds from Bird Island foraged in an area with densities of sardines and anchovies of 17.0 and 15.0 g m–2, respectively. Gannets from Malgas Island had to fly east of Cape Point to feed on sardines or anchovies, corresponding to a round trip of ≥400 km. In contrast, gannets from Bird Island were able to target natural prey much closer to their colony and took advantage of a rich food environment, at least during the study period.

Seabird energetics Diet sampled during the study was dominated by the deep-water hake Merluccius paradoxus for Malgas birds (78% by mass), and by sardines for birds from Bird Island (73% by mass, Fig. 4). Hake consumed by Cape gannets from Malgas occurs in waters 200 to 1000 m deep and is the primary target species of a demersal trawl fishery (Fairweather et al. 2006). The hakes taken by gannets are scavenged from behind trawlers. Sardines, by comparison, are epipelagic fish, which represent natural, live prey for Cape gannets (Berruti et al. 1993). The bioenergetics model (Enstipp et al. 2006) estimated the daily metabolism of Cape gannets breeding on Malgas to be 3.21 × BMR, with a necessary daily food intake of 1250 g of fish for the parent alone and 1415 g for 1 parent and its chick. On Bird Island, gannets had an estimated daily metabolism of 2.88 × BMR, with a predicted daily food requirement of 620 g of fish for self-maintenance and 710 g when including the chick’s needs. These results emphasise the energetic consequences of higher foraging effort in birds from Malgas and the lower calorific value of their prey compared to those from Bird Island; hakes have a calorific value of 4.07 kJ g–1 and sardines 8.59 kJ g–1 (Batchelor & Ross 1984). Breeding birds from Malgas Island have a daily food demand of > 50% of their body mass.

1%

7%

7%

3%

17%

92%

73%

Anchovy Sardine Hake

133

DISCUSSION Our study shows that Cape gannets are able to adapt their foraging effort according to food availability within the marine environment. Facing a scarcity of prey, birds increased their foraging effort but also turned to scavenging behind trawlers, taking prey of lower energy content. They were forced to invest more energy in reproduction, which was deemed hardly sustainable by our bioenergetics model. We suggest that low foraging efficiency is one of the factors underlying the decline in numbers of Cape gannets breeding on Malgas Island (Crawford et al. 2007). Our study confirms that estimates of foraging effort during the breeding season provide a powerful index of population health (Grémillet et al. 2006). The acoustic survey located very few anchovies and sardines off the west coast of South Africa during our study, with most sardines located in Nelson Mandela Bay (Fig. 3). Birds from the 2 colonies thus exploited contrasting foraging environments, with birds from Malgas facing a scarcity of natural prey, while the environment in Nelson Mandela Bay was more profitable. In a rich environment, birds from Bird Island could afford to spend more time resting at sea and flew less. The lack of a significant regression between foraging path length and foraging trip duration among Bird Island birds is unique compared to results found for other colonies of Cape and Northern gannets (Hamer et al. 2001, Lewis et al. 2006). By comparison, with few pelagic fish available, birds from Malgas were forced to increase their foraging effort substantially (Fig 3). They stayed longer at sea, travelled further, and spent longer periods flying per hour at sea than gannets from Bird Island, increasing their total foraging range 5-fold (Table 1). Birds from Malgas spent more time searching for their preferred prey. This finding is supported by the stronger regression between the amount of food returned to the colony and the time spent at sea for Malgas birds. However, they resorted to feeding behind trawlers (predominance of shallow dives; Table 1), bringing back fishery discards to the colony (Fig. 4). Finally, the greater variability in foraging trip duration of birds from Malgas Island indicated greater uncertainty in their foraging environment (Table 1), which accords with the high degree of variability of the overall Benguela system higher up in the food web (van der Lingen et al. 2006).

Other

Foraging in an unpredictable environment

Malgas Island

Bird Island

Fig. 4. Morus capensis. Percentage contribution by mass of prey items in the birds’ diet (n = 28 birds from each colony)

The low abundance of sardines and anchovies off the west coast during spring 2005 (Fig. 3) followed recent changes in the distribution of small pelagic fish, partic-

134

Mar Ecol Prog Ser 350: 127–136, 2007

ularly sardines, off South Africa, with a progressive southern and eastward shift since the mid-1990s (van der Lingen et al. 2005). The distribution of anchovy spawners on the Agulhas Bank has also shown an eastward shift, initiated in 1996, with the bulk of spawners observed during acoustic surveys now being found east of Cape Agulhas, whereas most were west of Cape Agulhas before 1996 (van der Lingen et al. 2002). The timing of this progressive eastward shift of both small pelagic fish species in the mid to late 1990s coincides with the start of the decrease in gannet population size on Malgas Island (Crawford et al. 2007), strongly suggesting that the 2 processes are linked. Moreover, decadal variation in the size of the Malgas Island breeding population (Crawford et al. 2007) can be explained by the fluctuations of anchovy and sardine populations (Schwartzlose et al. 1999), supporting the assumption that population growth is driven in part by food availability during the breeding season (Lewis et al. 2006). These fluctuations of small pelagics could be of great importance in marine ecosystems where these fish are predominant prey species, such as in the Benguela (Lluch-Belda et al. 1989), with known repercussions on top predators that depend on those prey, including Cape gannets (Crawford 1999).

Impact of fisheries on seabirds The Cape gannet diet reflects environmental variation in the abundance of sardines and anchovies (Berruti et al. 1993, Crawford 1999). In the early 2000s, the proportions of both species in gannet diet at west coast colonies decreased, with a concomitant increase in offal discarded from bottom trawlers (Marine and Coastal Management unpubl. data). The relative proportion of fishery wastes in gannet diet is not correlated with the amount of hake caught by the demersal fishery (Berruti et al. 1993), suggesting that hake is not a preferred prey item. Such prey-switching behaviour is one way in which seabirds may cope with environmental variability (Crawford 1999), but hake is of low calorific value compared to anchovies and sardines (Batchelor & Ross 1984). Gannets from Malgas Island were able to compensate for this to some extent by bringing back larger loads to the nest, resulting in meals with energetic values similar to those of the loads brought back by birds from Bird Island (Table 1). However, they were probably handicapped by the larger loads they were forced to carry. Moreover, previous studies have shown that feeding nestlings on a lipid-poor diet negatively impacts the fledglings’ body condition and cognitive abilities (Batchelor & Ross 1984, Kitaysky et al. 2005), which culminates in a higher mortality rate. A regime switch from lipid-rich

to lipid-poor prey can be responsible for changes in breeding success and population structure of piscivorous seabirds (the ‘junk-food hypothesis’, Anderson & Piatt 1999, Wanless et al. 2005). Fishery discards have become the main resource for some populations of breeding seabirds when natural prey stocks are depleted (Garthe et al. 1996, Votier et al. 2004). It has been hypothesised that increased fishing pressure in the 20th century and the associated increase in waste availability may have changed the structure and abundance of certain seabird communities (Abrams 1985), and it is known that such artificial food resources can affect foraging parameters and/or reproductive characteristics of birds (Arcos & Oro 1996, Oro et al. 2004). Future changes in fisheries management policies, such as increased use of offal for industrial purposes and reducing the amount of nontarget species being discarded, may further threaten seabird populations reliant on discards (Furness 2003, Votier et al. 2004). Although some fisheries increase food availability for seabirds through waste, other fisheries compete directly with seabirds when harvesting their main natural prey (Furness 2003). In the southern Benguela ecosystem, the sardines and anchovies that are the main target of the pelagic fishery (van der Lingen et al. 2006) also are the main food resource of several endemic seabirds, including Cape gannets, the African penguin Spheniscus demersus, and the Cape cormorant Phalacrocorax capensis (Crawford 1999). Obviously, the negative impact of competition between fisheries and seabirds is enhanced when prey becomes scarcer, and a substantial negative impact of fisheries on Cape gannets in the northern Benguela (off Namibia) has been reported previously (Crawford 1999). In the 1950s and 1960s, heavy fishing pressure on sardines without appropriate management policies led to a collapse of 85% of the Namibian populations of gannets, which meant an overall decrease of 40% in the entire Cape gannet population (Crawford et al. 2007).

Conservation implications Although life-history traits of seabirds act to buffer populations against short-term fluctuations in their food supply, Cape gannets do not seem to be able to cope with recent decadal scale environmental changes, and the overall population of Cape gannets is decreasing. On Malgas Island (which until recently supported 40% of the world population of gannets), the gannet population has started to show a dramatic decrease, arising both from reduced prey availability and increased predation by the Cape fur seals Arctocephalus pusillus, which may kill up to 80% of fledg-

Pichegru et al.: Seabird foraging on live prey versus fishery discards

lings (Makhado et al. 2006). Great white pelicans (Pelecanus onocrotalus) have also recently started feeding on gannet chicks on Malgas Island (L. Pichegru pers. obs.). Together, these threats weigh heavily on a bird already classified as vulnerable (see the 2006 IUCN Red List of Threatened Species, available at www.iucnredlist.org) with a breeding area restricted to only 6 colonies, 5 of which are declining (Crawford et al. 2007). Actions to mitigate these threats are required urgently. Acknowledgements. This study was funded by the Centre National de la Recherche National via an ACI jeunes chercheuses et jeunes chercheurs to D.G. and by the DEPEIPHC-CNRS. L.P. was funded by a studentship of the French Ministry of Research. We thank South African National Parks for granting access to protected areas and the University of Cape Town for transport and logistic support. Marine and Coastal Management kindly provided gannet diet data and logistical support in the field. A. Kato at the National Institute of Polar Research kindly lent us the data loggers and M. Enstipp adapted his routine of calculation of the energetic requirements for breeding Cape gannets. R. Mullers, C. Le Bohec, V. Viera, S. Sari, and D. Chevallier helped at different stages of this work, and the comments of 3 anonymous referees greatly improved our manuscript. We warmly thank them all.

LITERATURE CITED Abrams RW (1985) Pelagic seabird community structure in the Southern Benguela region: changes in response to man’s activities? Biol Conserv 32:33–49 Adams NJ, Klages NTW (1999) Foraging effort and prey choice in Cape gannets. S Afr J Mar Sci 21:157–163 Anderson PJ, Piatt JF (1999) Community reorganization in the Gulf of Alaska following ocean climate regime shift. Mar Ecol Prog Ser 189:117–123 Arcos JM, Oro D (1996) Changes in foraging range of Audouin’s gulls Larus audouinii in relation to a trawler moratorium in the Western Mediterranean. Colonial Waterbirds 19:128–131 Bakun A (1996) Patterns in the ocean: ocean processes and marine population dynamics. California Sea Grant College System, National Oceanic and Atmospheric Administration (NOAA), La Jolla, CA Barange M, Hamptom I, Roel BA (1999) Trends in the abundance and distribution of anchovy and sardine on the South African continental shelf in the 1990s, deduced from acoustic surveys. S Afr J Mar Sci 21:367–391 Batchelor AL, Ross GJB (1984) The diet and implications of dietary change of Cape gannets on Bird Island, Nelson Mandela Bay. Ostrich 55:45–63 Berruti A, Underhill LG, Shelton PA, Moloney CL, Crawford RJM (1993) Seasonal and interannual variation in the diet of two colonies of the Cape gannet (Morus capensis) between 1977–78 and 1989. Colon Waterbirds 16:158–175 Crawford RJM (1999) Seabird responses to long-term changes of prey resources off southern Africa. In: Adams NJ, Slotow RH (eds) Proc 22nd Int Ornithol Congr, Durban. BirdLife South Africa, Johannesburg, p 688–705 Crawford RJM (2005) Cape gannet. In: Hockey PAR, Dean WRJ, Ryan PG (eds) Roberts’ birds of Southern Africa (7th edn). John Voelcker Bird Book Fund, Cape Town, p 565–567

135

Crawford RJM, Shannon LV, Pollock DE (1987) The Benguela ecosystem. 4. The major fish and invertebrate resources. In: Barnes M (ed) Oceanography and marine biology: an annual review, Vol 25. Aberdeen University Press, p 353–505 Crawford RJM, Ryan P, Williams AJ (1991) Seabird consumption and production in the Benguela and Western Agulhas ecosystems. S Afr J Mar Sci 11:357–375 Crawford RJM, Dundee BL, Dyer BM, Klages NTW, Meÿer MA, Upfold L (2007) Trends in numbers of Cape gannets (Morus capensis) 1956/57-2005/06, with a consideration of the influence of food and other factors. ICES J Mar Sci 64:169–177 Enstipp MR, Daunt F, Wanless S, Humphreys EM, Hamer KC, Benvenuti S, Grémillet D (2006) Foraging energetics of North Sea birds confronted with fluctuating prey availability. In: Boyd IL, Wanless S, Camphuysen CJ (eds) Top predators in marine ecosystems (symposium Zool Soc Lond). Cambridge University Press, Cambridge, p 191–210 Fairweather TP, Booth AJ, Sauer WHH, Leslie RW (2006) Spatial description of hake-directed fishing activity off the west coast of South Africa. Afr J Mar Sci 28:13–24 Furness RW (2003) Impacts of fisheries on seabird communities. Sci Mar 67:33–45 Furness RW, Tasker ML (2000) Seabird-fishery interactions: quantifying the sensitivity of seabirds to reductions in sandeel abundance, and identification of key areas for sensitive seabirds in the North Sea. Mar Ecol Prog Ser 202: 253–264 Garthe S, Camphuysen CJ, Furness RW (1996) Amounts of discards by commercial fisheries and their significance as food for seabirds in the North Sea. Mar Ecol Prog Ser 136: 1–11 Girard I, Ouellet JP, Courtois R, Dussault C, Breton L (2002) Effects of sampling effort based on GPS telemetry on homerange size estimations. J Wildl Manag 66:1290–1300 Grémillet D, Kuntz G, Delbart F, Mellet M and 6 others (2004a) Linking the foraging performance of a marine predator to local prey abundance. Funct Ecol 18:1–7 Grémillet D, Dell’Omo G, Ryan PG, Peters G, Ropert-Coudert Y, Weeks S (2004b) Offshore diplomacy, or how seabirds mitigate intra-specific competition: a case study based on GPS tracking of cape gannets from neighbouring breeding sites. Mar Ecol Prog Ser 268:265–279 Grémillet D, Pichegru L, Siorat F, Georges JY (2006) Conservation implications of the apparent mismatch between population dynamics and foraging effort in Northern gannets from the English Channel. Mar Ecol Prog Ser 319: 15–25 Griffiths CL, van Sittert L, Best PB, Brown AC and 14 others (2004) Impacts of human activities on marine life in the Benguela: a historical overview. In: Gibson RN et al. (eds) Oceanography and marine biology: an annual review, Vol 42. CRC Press, Boca Raton, FL, p 303–392 Hamer KC, Phillips RA, Hill JK, Wanless S, Wood AG (2001) Contrasting foraging strategies of gannets Morus bassanus at two North Atlantic colonies: foraging trip duration and foraging area fidelity. Mar Ecol Prog Ser 224:283–290 Kitaysky AS, Kitaiskaia EV, Piatt JF, Wingfield JC (2005) A mechanistic link between chick diet and decline in seabirds? Proc R Soc B 273:445–450 Klages NTW, Willis AB, Ross GJB (1992) Variability in the diet of the Cape gannet at Bird Island, Nelson Mandela Bay, South Africa. S Afr J Mar Sci 12:761–771 Lescroël A, Bost CA (2005) Foraging under contrasted oceanographic conditions: the Gentoo penguin at Kergue-

136

Mar Ecol Prog Ser 350: 127–136, 2007

len Archipelago. Mar Ecol Prog Ser 302:245–261 Lewis S, Grémillet D, Daunt F, Ryan PG, Crawford RJM, Wanless S (2006) Using behavioural variables to identify proximate causes of population change in a seabird. Oecologia 147:606–614 Lluch-Belda D, Crawford RJM, Kawasaki T, MacCall A, Parrish RH, Schwartzlose RA, Smith PE (1989) World-wide fluctuations of sardine and anchovy stocks: the regime problem. S Afr J Mar Sci 8:195–205 Makhado AB, Crawford RJM, Underhill LG (2006) Impact of predation by Cape fur seals Arctocephalus pusillus pusillus on Cape gannets Morus capensis at Malgas Island, Western Cape, South Africa. Afr J Mar Sci 28(3/4): 681–687 Miller DCM, Moloney CL, van der Lingen CD, Lett C, Mullon C, Field JG (2006) Modelling the effects of physical-biological interactions and spatial variability in spawning and nursery areas on transport and retention of sardine Sardinops sagax eggs and larvae in the southern Benuela ecosystem. J Mar Syst 61:212–229 Oro D, Cam E, Pradel R, Martínez-Abraín A (2004) Influence of food availability on demography and local population dynamics in a long-lived seabird. Proc R Soc B 271: 387–396 Ropert-Coudert Y, Grémillet D, Kato A, Ryan P, Naito Y, Le Maho Y (2004) A fine-scale time budget of Cape gannets provides insights into the foraging strategies of coastal seabirds. Anim Behav 67:985–992 Ryan PG, Petersen S, Peters G, Grémillet D (2004) GPS tracking a marine predator: the effects of precision, resolution and sampling rate on foraging tracks of African Penguins. Mar Biol 145:215–223 Schreiber EA, Burger J (2002) Seabirds and the marine envi-

ronment. In: Schreiber EA, Burger J (eds) Biology of marine birds. CRC Press, Boca Raton, FL, p 87–113 Schwartzlose RA, Alheit J, Bakun A, Baumgartner TR and 17 others (1999) Worldwide large-scale fluctuations of sardine and anchovy populations. S Afr J Mar Sci 21:289–347 Shannon LV, O’Toole MJ (2003) Sustainability of the Benguela: ex Africa semper aliquid novi. In: Hempel G, Sherman K (eds) Large marine ecosystem: trends in exploitation protection and research. Elsevier, Amsterdam, p 227–253 van der Lingen CD, Coetzee J, Hutchings L (2002) Temporal shifts in the spatial distribution of anchovy spawners and their eggs in the Southern Benguela: implications for recruitment. GLOBEC Rep 16:46–48 van der Lingen CD, Coetzee J, Demarcq H, Drapeau L, Fairweather TP, Hutchings L (2005) An eastward shift in the distribution of the southern Benguela sardine. GLOBEC Newsl 11:17–22 van der Lingen CD, Shannon LJ, Cury P, Kreiner A, Moloney CL, Roux JP, Vaz-Velo F (2006) Resource and ecosystem variability, including regime shifts, in the Benguela Current system. In: Shannon V, Hempel G, Malanotte-Rizzoli P, Moloney CL, Woods J (eds) Benguela: predicting a large marine ecosystem. Large marine ecosystem series, Vol 14. Elsevier, Amsterdam, p 147–184 Votier SC, Furness RW, Bearhop S, Crane JE and 11 others (2004) Changes in fisheries discard rates and seabird communities. Nature 427:727–730 Walmsley SA, Leslie RW, Sauer WHH (2007) Managing South Africa’s trawl bycatch. ICES J Mar Sci 64:405–412 Wanless S, Harris MP, Redman P, Speakman JR (2005) Low energy value of fish as a probable cause of a major seabird breeding failure in the North Sea. Mar Ecol Prog Ser 294:1–8

Editorial responsibility: Otto Kinne (Editor-in-Chief), Oldendorf/Luhe, Germany

Submitted: December 19, 2006; Accepted: June 25, 2007 Proofs received from author(s): November 2, 2007